Ephedra californica recovery
Purpose
To explore Ephedra californica recovery following clipping. Year two in a long-term shrub removal experiment. Aboveground clipping. Panoche Hills Ecological Reserve.
Removal done by A. Liczner, A. Filazzola, T. Noble, and M. Westphal. Loss of shrub effects on animals experiment separate from this survey initiated by A. Liczner and completed 2016. Eva Gruber surveyed for resprouting.
Ephedra foundation effects
An examination of the evidence associated with pattern analyses between ephedra and other plant species at Panoche Hills Ecological Reserve.
#Data
#Primary
foundation <- read_csv("data/ephedra.foundation.csv")
foundation <- foundation %>% mutate(volume = ((length + width)/2)^3*3.14*(1/3))
foundation$year <- as.character(foundation$year)
foundation$site <- as.character(foundation$site)
foundation## # A tibble: 450 × 11
## year region site rep microsite length width height abundance
## <chr> <chr> <chr> <int> <chr> <dbl> <dbl> <dbl> <int>
## 1 2016 Panoche 1.3 1 Shrub 3.77 2.38 1.59 263
## 2 2016 Panoche 1.3 1 Open NA NA NA 106
## 3 2016 Panoche 1.3 2 Shrub 3.22 2.38 1.84 276
## 4 2016 Panoche 1.3 2 Open NA NA NA 245
## 5 2016 Panoche 1.3 3 Shrub 1.95 1.30 0.59 217
## 6 2016 Panoche 1.3 3 Open NA NA NA 175
## 7 2016 Panoche 1.3 4 Shrub 4.21 4.39 1.71 283
## 8 2016 Panoche 1.3 4 Open NA NA NA 214
## 9 2016 Panoche 1.3 5 Shrub 3.41 3.17 1.55 241
## 10 2016 Panoche 1.3 5 Open NA NA NA 233
## # ... with 440 more rows, and 2 more variables: richness <int>,
## # volume <dbl>#Effect sizes
foundation.rii <- read_csv("data/ephedra.foundation.rii.csv")
foundation.rii <- foundation.rii %>% mutate(volume = ((length + width)/2)^3*3.14*(1/3))
foundation.rii$year <- as.character(foundation.rii$year)
foundation.rii$site <- as.character(foundation.rii$site)
foundation.rii <- foundation.rii %>% filter(abundance < 0.8 & abundance > -0.7)
foundation.rii## # A tibble: 223 × 10
## year region site rep microsite length width abundance richness
## <chr> <chr> <chr> <int> <chr> <dbl> <dbl> <dbl> <dbl>
## 1 2013 Panoche 1.1 1 Shrub 4.5 4.8 0.22580645 0.0000000
## 2 2013 Panoche 1.1 2 Shrub 2.9 2.1 0.09803922 0.1111111
## 3 2013 Panoche 1.1 3 Shrub 1.2 1.2 0.04761905 0.0000000
## 4 2013 Panoche 1.1 4 Shrub 1.9 2.1 0.21739130 0.0000000
## 5 2013 Panoche 1.1 5 Shrub 3.4 4.4 0.16666667 -0.1428571
## 6 2013 Panoche 1.1 6 Shrub 4.1 2.1 0.39622641 -0.1428571
## 7 2013 Panoche 1.1 7 Shrub 2.4 2.3 0.21212121 0.1428571
## 8 2013 Panoche 1.1 8 Shrub 3.2 3.0 0.16666667 0.0000000
## 9 2013 Panoche 1.1 9 Shrub 4.1 3.8 0.35849057 -0.1428571
## 10 2013 Panoche 1.1 10 Shrub 3.2 3.1 0.31343284 -0.1428571
## # ... with 213 more rows, and 1 more variables: volume <dbl>rii.2013 <- foundation.rii %>% filter(year == "2013")
rii.2014 <- foundation.rii %>% filter(year == "2014")
rii.2015 <- foundation.rii %>% filter(year == "2015")
rii.2016 <- foundation.rii %>% filter(year == "2016")
#Rii
#differences from open
ggplot(foundation.rii, aes(year, abundance)) + geom_boxplot(fill = "darkseagreen3") + ylab("Rii abundance")
ggplot(foundation.rii, aes(year,richness)) + geom_boxplot(fill = "darkseagreen3") + ylab("Rii richness")
#models for different between years
require(lmerTest)
m <- lmer(abundance~(1|year) + volume, data = foundation.rii)
summary(m) ## Linear mixed model fit by REML t-tests use Satterthwaite approximations
## to degrees of freedom [lmerMod]
## Formula: abundance ~ (1 | year) + volume
## Data: foundation.rii
##
## REML criterion at convergence: 70.2
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.88876 -0.58976 0.00998 0.71781 2.19688
##
## Random effects:
## Groups Name Variance Std.Dev.
## year (Intercept) 0.003714 0.06094
## Residual 0.072891 0.26998
## Number of obs: 223, groups: year, 4
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 0.200176 0.039885 4.360000 5.019 0.0059 **
## volume -0.001036 0.000522 220.410000 -1.984 0.0485 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## volume -0.443#anova(m, test = "Chisq")
m <- lmer(richness~(1|year) + volume, data = foundation.rii)
summary(m) ## Linear mixed model fit by REML t-tests use Satterthwaite approximations
## to degrees of freedom [lmerMod]
## Formula: richness ~ (1 | year) + volume
## Data: foundation.rii
##
## REML criterion at convergence: 33.9
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.9255 -0.8477 0.0826 0.6685 2.7652
##
## Random effects:
## Groups Name Variance Std.Dev.
## year (Intercept) 0.02499 0.1581
## Residual 0.06033 0.2456
## Number of obs: 223, groups: year, 4
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) -1.110e-01 8.247e-02 3.270e+00 -1.346 0.2640
## volume 8.235e-04 4.765e-04 2.186e+02 1.728 0.0853 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## volume -0.195#anova(m, test = "Chisq")
#different from zero
#ab
t1 <- t.test(foundation.rii$abundance, mu = 0)
t1##
## One Sample t-test
##
## data: foundation.rii$abundance
## t = 8.5519, df = 222, p-value = 1.988e-15
## alternative hypothesis: true mean is not equal to 0
## 95 percent confidence interval:
## 0.1215495 0.1943441
## sample estimates:
## mean of x
## 0.1579468#richness
t2 <- t.test(foundation.rii$richness, mu = 0)
t2##
## One Sample t-test
##
## data: foundation.rii$richness
## t = -6.1717, df = 222, p-value = 3.171e-09
## alternative hypothesis: true mean is not equal to 0
## 95 percent confidence interval:
## -0.15553974 -0.08024948
## sample estimates:
## mean of x
## -0.1178946# ab each year
t1 <- t.test(rii.2013$abundance, mu = 0)
t1##
## One Sample t-test
##
## data: rii.2013$abundance
## t = 8.6815, df = 43, p-value = 5.253e-11
## alternative hypothesis: true mean is not equal to 0
## 95 percent confidence interval:
## 0.1640679 0.2633583
## sample estimates:
## mean of x
## 0.2137131t1 <- t.test(rii.2014$abundance, mu = 0)
t1##
## One Sample t-test
##
## data: rii.2014$abundance
## t = 4.4799, df = 44, p-value = 5.256e-05
## alternative hypothesis: true mean is not equal to 0
## 95 percent confidence interval:
## 0.1292521 0.3406396
## sample estimates:
## mean of x
## 0.2349459t1 <- t.test(rii.2015$abundance, mu = 0)
t1##
## One Sample t-test
##
## data: rii.2015$abundance
## t = 2.9259, df = 44, p-value = 0.005416
## alternative hypothesis: true mean is not equal to 0
## 95 percent confidence interval:
## 0.03062158 0.16618316
## sample estimates:
## mean of x
## 0.09840237t1 <- t.test(rii.2016$abundance, mu = 0)
t1##
## One Sample t-test
##
## data: rii.2016$abundance
## t = 3.9614, df = 88, p-value = 0.000151
## alternative hypothesis: true mean is not equal to 0
## 95 percent confidence interval:
## 0.06057307 0.18252997
## sample estimates:
## mean of x
## 0.1215515#richness by year
t2 <- t.test(rii.2013$richness, mu = 0)
t2##
## One Sample t-test
##
## data: rii.2013$richness
## t = 0.075117, df = 43, p-value = 0.9405
## alternative hypothesis: true mean is not equal to 0
## 95 percent confidence interval:
## -0.03475463 0.03744386
## sample estimates:
## mean of x
## 0.001344615t2 <- t.test(rii.2014$richness, mu = 0)
t2##
## One Sample t-test
##
## data: rii.2014$richness
## t = 2.668, df = 44, p-value = 0.01064
## alternative hypothesis: true mean is not equal to 0
## 95 percent confidence interval:
## 0.02465005 0.17688174
## sample estimates:
## mean of x
## 0.1007659t2 <- t.test(rii.2015$richness, mu = 0)
t2##
## One Sample t-test
##
## data: rii.2015$richness
## t = -4.6118, df = 44, p-value = 3.435e-05
## alternative hypothesis: true mean is not equal to 0
## 95 percent confidence interval:
## -0.24208568 -0.09484554
## sample estimates:
## mean of x
## -0.1684656t2 <- t.test(rii.2016$richness, mu = 0)
t2##
## One Sample t-test
##
## data: rii.2016$richness
## t = -8.6023, df = 88, p-value = 2.702e-13
## alternative hypothesis: true mean is not equal to 0
## 95 percent confidence interval:
## -0.3223218 -0.2013451
## sample estimates:
## mean of x
## -0.2618334#volume
ggplot(foundation.rii, aes(volume, abundance)) + geom_point(aes(color = year)) + geom_smooth(method = "lm")
ggplot(foundation.rii, aes(volume, abundance, color = year)) + geom_point() + geom_smooth(method = "lm", se = FALSE)
ggplot(foundation.rii, aes(volume, richness)) + geom_point(aes(color=year)) + geom_smooth(method = "lm")
ggplot(foundation.rii, aes(volume, richness, color = year)) + geom_point() + geom_smooth(method = "lm", se = FALSE)
#posthoc tests from main models
m <- lm(abundance~volume, data = foundation.rii)
m##
## Call:
## lm(formula = abundance ~ volume, data = foundation.rii)
##
## Coefficients:
## (Intercept) volume
## 0.1889542 -0.0009065summary(m)##
## Call:
## lm(formula = abundance ~ volume, data = foundation.rii)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.80630 -0.16057 0.01776 0.18475 0.64730
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.1889542 0.0257240 7.345 3.91e-12 ***
## volume -0.0009065 0.0005259 -1.724 0.0862 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.2746 on 221 degrees of freedom
## Multiple R-squared: 0.01327, Adjusted R-squared: 0.0088
## F-statistic: 2.971 on 1 and 221 DF, p-value: 0.08617m <- lm(richness~volume, data = foundation.rii)
m##
## Call:
## lm(formula = richness ~ volume, data = foundation.rii)
##
## Coefficients:
## (Intercept) volume
## -0.1515101 0.0009828summary(m)##
## Call:
## lm(formula = richness ~ volume, data = foundation.rii)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.54197 -0.20375 0.00403 0.18907 0.79721
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -0.1515101 0.0265882 -5.698 3.84e-08 ***
## volume 0.0009828 0.0005436 1.808 0.072 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.2838 on 221 degrees of freedom
## Multiple R-squared: 0.01457, Adjusted R-squared: 0.01011
## F-statistic: 3.268 on 1 and 221 DF, p-value: 0.07198#test 2016 only
m <- lm(abundance~volume, data = rii.2016)
m##
## Call:
## lm(formula = abundance ~ volume, data = rii.2016)
##
## Coefficients:
## (Intercept) volume
## 0.217302 -0.002641summary(m)##
## Call:
## lm(formula = abundance ~ volume, data = rii.2016)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.78160 -0.12944 0.00135 0.23578 0.46049
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.217302 0.040966 5.304 8.49e-07 ***
## volume -0.002641 0.000796 -3.318 0.00133 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.2743 on 87 degrees of freedom
## Multiple R-squared: 0.1123, Adjusted R-squared: 0.1021
## F-statistic: 11.01 on 1 and 87 DF, p-value: 0.001325m <- lm(richness~volume, data = rii.2016)
m##
## Call:
## lm(formula = richness ~ volume, data = rii.2016)
##
## Coefficients:
## (Intercept) volume
## -0.344154 0.002271summary(m)##
## Call:
## lm(formula = richness ~ volume, data = rii.2016)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.40545 -0.26604 0.03897 0.21848 0.59324
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -0.344154 0.041271 -8.339 1.01e-12 ***
## volume 0.002271 0.000802 2.831 0.00576 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.2763 on 87 degrees of freedom
## Multiple R-squared: 0.08437, Adjusted R-squared: 0.07385
## F-statistic: 8.017 on 1 and 87 DF, p-value: 0.005756#Primary data
#ggplot(foundation, aes(microsite, abundance)) + geom_boxplot() + facet_grid(~year)
#ggplot(foundation, aes(microsite, richness)) + geom_boxplot() + facet_grid(~year)
#ggplot(foundation, aes(volume, abundance, color = year)) + geom_point() + geom_smooth(method = "lm")
#ggplot(foundation, aes(volume, richness, color = year)) + geom_point() + geom_smooth(method = "lm")Recovery and stem data
#recovery
data <- read_csv("data/EH.recovery.2016.csv")
data <- data %>% mutate(volume = ((length + width)/2)^3*3.14*(1/3)) %>% arrange(desc(resprout))
data## # A tibble: 40 × 9
## date census ID length width height resprout
## <chr> <chr> <int> <dbl> <dbl> <dbl> <int>
## 1 nov.2016 post.2 182 NA NA NA 15
## 2 nov.2016 post.2 217 1.15 1.25 0.85 15
## 3 nov.2016 post.2 303 3.50 2.80 0.60 13
## 4 nov.2016 post.2 614 1.70 1.15 0.43 12
## 5 nov.2016 post.2 146 3.40 2.15 0.70 10
## 6 nov.2016 post.2 621 1.85 0.90 0.65 10
## 7 nov.2016 post.2 651 2.55 2.40 1.00 10
## 8 nov.2016 post.2 137 1.90 3.30 0.60 9
## 9 nov.2016 post.2 245 3.20 2.50 0.80 8
## 10 nov.2016 post.2 271 2.80 2.55 0.71 8
## # ... with 30 more rows, and 2 more variables: observations <chr>,
## # volume <dbl>#stem data
wet.weights <-read.csv("data/wet.weights.csv")
total.weight <- wet.weights %>% select(ID, total.weight)
stems <- read_csv("data/stem.analyses.csv")EDA field data & wet weight
#current volume used####
#data viz
nov.2016 <- data %>% filter(date == "nov.2016") #only resurvey
nov.2016 <- left_join(nov.2016, total.weight, by = "ID")
p1 <- ggplot(nov.2016, aes(resprout, weight= volume)) +
geom_histogram(binwidth = 2, fill = "dodgerblue") +
xlab("number of shoots resprouted") +
ylab("relative weighted frequency by current volume")
ggplotly(p1)p1.1 <- ggplot(nov.2016, aes(ID, resprout)) +
geom_point(aes(size = volume), color = "dodgerblue")
ggplotly(p1.1)nov.2016 <- nov.2016 %>% filter(resprout < 15) #filter out my post hoc classifications
p2 <- ggplot(nov.2016, aes(volume, resprout)) +
geom_point(color = "dodgerblue") +
xlab("current shrub volume") +
ylab("number of resprouted shoots") +
geom_smooth(method = lm)
ggplotly(p2)p3 <- ggplot(nov.2016, aes(volume, total.weight)) +
geom_point(color = "dodgerblue") +
xlab("current shrub volume") +
ylab("total wet weight") +
geom_smooth(method = lm)
ggplotly(p3)#number of resprouted shoots predict wet weight
p4 <- ggplot(nov.2016, aes(resprout, total.weight)) +
geom_point(color = "dodgerblue") +
xlab("resprouted shoot numbers") +
ylab("total wet weight") +
geom_smooth(method = lm)
ggplotly(p4)#quick test
fit <- lm(total.weight ~ resprout, data= nov.2016)
summary(fit)##
## Call:
## lm(formula = total.weight ~ resprout, data = nov.2016)
##
## Residuals:
## Min 1Q Median 3Q Max
## -2127.08 -615.68 -99.73 1120.22 2763.02
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 594.48 738.92 0.805 0.4337
## resprout 189.05 95.54 1.979 0.0665 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1390 on 15 degrees of freedom
## (1 observation deleted due to missingness)
## Multiple R-squared: 0.207, Adjusted R-squared: 0.1541
## F-statistic: 3.916 on 1 and 15 DF, p-value: 0.06649#simple main model
m1 <- glm(resprout ~ volume, data = nov.2016)
summary(m1)##
## Call:
## glm(formula = resprout ~ volume, data = nov.2016)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -3.9476 -1.9340 -0.7416 0.9654 7.2943
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.94757 0.97041 4.068 0.000895 ***
## volume 0.25031 0.06696 3.738 0.001793 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 8.557791)
##
## Null deviance: 256.50 on 17 degrees of freedom
## Residual deviance: 136.92 on 16 degrees of freedom
## AIC: 93.605
##
## Number of Fisher Scoring iterations: 2anova(m1, test = "Chisq")## Analysis of Deviance Table
##
## Model: gaussian, link: identity
##
## Response: resprout
##
## Terms added sequentially (first to last)
##
##
## Df Deviance Resid. Df Resid. Dev Pr(>Chi)
## NULL 17 256.50
## volume 1 119.58 16 136.93 0.0001855 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1m1.1 <- glm(total.weight ~ volume, data = nov.2016)
summary(m1.1)##
## Call:
## glm(formula = total.weight ~ volume, data = nov.2016)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2045.56 -437.74 -66.36 467.51 2879.01
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 887.75 414.97 2.139 0.04926 *
## volume 93.35 27.83 3.354 0.00435 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 1392680)
##
## Null deviance: 36559960 on 16 degrees of freedom
## Residual deviance: 20890205 on 15 degrees of freedom
## (1 observation deleted due to missingness)
## AIC: 292.61
##
## Number of Fisher Scoring iterations: 2anova(m1.1, test = "Chisq")## Analysis of Deviance Table
##
## Model: gaussian, link: identity
##
## Response: total.weight
##
## Terms added sequentially (first to last)
##
##
## Df Deviance Resid. Df Resid. Dev Pr(>Chi)
## NULL 16 36559960
## volume 1 15669755 15 20890205 0.0007956 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1#setup volume####
setup <- data %>% filter(date == "april.2013") %>%
select(ID, volume) %>% rename(setup.vol = volume)
merged.data <- left_join(nov.2016, setup)
#repeat all above worflow code
p3 <- ggplot(merged.data, aes(resprout, weight= setup.vol)) +
geom_histogram(binwidth = 2, fill = "dodgerblue") +
xlab("number of shoots resprouted") +
ylab("relative weighted frequency by setup volume")
ggplotly(p3)p3.1 <- ggplot(merged.data, aes(total.weight, weight= setup.vol)) +
geom_histogram(binwidth = 1000, fill = "dodgerblue") +
xlab("total wet weight regrowth") +
ylab("relative weighted frequency by setup volume")
ggplotly(p3.1)p3.2 <- ggplot(merged.data, aes(ID, resprout)) +
geom_point(aes(size = setup.vol), color = "dodgerblue")
ggplotly(p3.2)merged.data <- merged.data %>% filter(resprout < 15) #filter out my post hoc classifications
p4 <- ggplot(merged.data, aes(setup.vol, resprout)) +
geom_point(color = "dodgerblue") +
xlab("setup shrub volume") +
ylab("number of resprouted shoots") +
geom_smooth(method = lm)
ggplotly(p4)p4.1 <- ggplot(merged.data, aes(setup.vol, total.weight)) +
geom_point(color = "dodgerblue") +
xlab("setup shrub volume") +
ylab("total wet weight regrowth") +
geom_smooth(method = lm)
ggplotly(p4.1)#simple model
m2 <- glm(resprout ~ setup.vol, data = merged.data)
summary(m2)##
## Call:
## glm(formula = resprout ~ setup.vol, data = merged.data)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -4.989 -2.966 -1.361 2.726 6.840
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.78551 1.40967 3.395 0.0037 **
## setup.vol 0.05216 0.03352 1.556 0.1392
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 13.92356)
##
## Null deviance: 256.50 on 17 degrees of freedom
## Residual deviance: 222.78 on 16 degrees of freedom
## AIC: 102.37
##
## Number of Fisher Scoring iterations: 2anova(m2, test = "Chisq")## Analysis of Deviance Table
##
## Model: gaussian, link: identity
##
## Response: resprout
##
## Terms added sequentially (first to last)
##
##
## Df Deviance Resid. Df Resid. Dev Pr(>Chi)
## NULL 17 256.50
## setup.vol 1 33.723 16 222.78 0.1196m2.1 <- glm(total.weight ~ setup.vol, data = merged.data)
summary(m2)##
## Call:
## glm(formula = resprout ~ setup.vol, data = merged.data)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -4.989 -2.966 -1.361 2.726 6.840
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.78551 1.40967 3.395 0.0037 **
## setup.vol 0.05216 0.03352 1.556 0.1392
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 13.92356)
##
## Null deviance: 256.50 on 17 degrees of freedom
## Residual deviance: 222.78 on 16 degrees of freedom
## AIC: 102.37
##
## Number of Fisher Scoring iterations: 2anova(m2.1, test = "Chisq")## Analysis of Deviance Table
##
## Model: gaussian, link: identity
##
## Response: total.weight
##
## Terms added sequentially (first to last)
##
##
## Df Deviance Resid. Df Resid. Dev Pr(>Chi)
## NULL 16 36559960
## setup.vol 1 698277 15 35861683 0.5889#Proportion regrowth by initial shrub size
merged.data <-merged.data %>% mutate(relative.regrowth = resprout/setup.vol) #calculate proportion reprouting by initial size
merged.data <-merged.data %>% mutate(relative.biomass.regrowth = total.weight/setup.vol) #calculate proportion reprouting by initial size
p5 <- ggplot(merged.data, aes(volume, relative.regrowth)) + geom_point(color = "dodgerblue") + xlab("current shrub volume") +
geom_smooth(method = lm) +
ylab("regrowth weighted by initial shrub size") +
scale_y_continuous(breaks=c(-2,-1,0,1,2))
ggplotly(p5)p5.1 <- ggplot(merged.data, aes(volume, relative.biomass.regrowth)) + geom_point(color = "dodgerblue") + xlab("current shrub volume") +
geom_smooth(method = lm) +
ylab("biomass regrowth weighted by initial shrub size") +
scale_y_continuous(breaks=c(-2,-1,0,1,2))
ggplotly(p5.1)m3 <- lm(merged.data$relative.regrowth ~ merged.data$volume)
summary(m3) # slope is not significantly different from 0 so no difference in proportionate regrowth##
## Call:
## lm(formula = merged.data$relative.regrowth ~ merged.data$volume)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.39398 -0.29158 -0.16772 -0.06767 1.91933
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.393979 0.206191 1.911 0.0741 .
## merged.data$volume 0.001213 0.014228 0.085 0.9331
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.6216 on 16 degrees of freedom
## Multiple R-squared: 0.0004539, Adjusted R-squared: -0.06202
## F-statistic: 0.007266 on 1 and 16 DF, p-value: 0.9331anova(m3, test = "Chisq")## Analysis of Variance Table
##
## Response: merged.data$relative.regrowth
## Df Sum Sq Mean Sq F value Pr(>F)
## merged.data$volume 1 0.0028 0.00281 0.0073 0.9331
## Residuals 16 6.1818 0.38636m3.1 <- lm(merged.data$relative.biomass.regrowth ~ merged.data$volume)
summary(m3.1) # slope is not significantly different from 0 so no difference in proportionate regrowth##
## Call:
## lm(formula = merged.data$relative.biomass.regrowth ~ merged.data$volume)
##
## Residuals:
## Min 1Q Median 3Q Max
## -153.97 -109.28 -97.07 -26.06 1070.12
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 122.584 103.349 1.186 0.254
## merged.data$volume 1.980 6.931 0.286 0.779
##
## Residual standard error: 293.9 on 15 degrees of freedom
## (1 observation deleted due to missingness)
## Multiple R-squared: 0.00541, Adjusted R-squared: -0.0609
## F-statistic: 0.0816 on 1 and 15 DF, p-value: 0.779anova(m3.1, test = "Chisq")## Analysis of Variance Table
##
## Response: merged.data$relative.biomass.regrowth
## Df Sum Sq Mean Sq F value Pr(>F)
## merged.data$volume 1 7049 7049 0.0816 0.779
## Residuals 15 1295773 86385EDA stems & dry weight
merged.data <- left_join(merged.data, stems, by = "ID")
#Mass EDA####
#Did wet predict dry weights?
p <- ggplot(merged.data, aes(total.weight, total.dry.weight.g)) +
geom_point(color = "dodgerblue") +
xlab("total wet weight") +
ylab("total dry weight") +
geom_smooth(method = lm)
ggplotly(p)fit <- lm(total.dry.weight.g ~ total.weight, data= merged.data)
summary(fit)##
## Call:
## lm(formula = total.dry.weight.g ~ total.weight, data = merged.data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -670.37 -23.23 37.01 37.01 2341.93
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -171.98974 49.50471 -3.474 0.000573 ***
## total.weight 0.95274 0.01173 81.232 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 438.2 on 370 degrees of freedom
## (172 observations deleted due to missingness)
## Multiple R-squared: 0.9469, Adjusted R-squared: 0.9468
## F-statistic: 6599 on 1 and 370 DF, p-value: < 2.2e-16#Volume by dry weight
p <- ggplot(merged.data, aes(volume, total.dry.weight.g)) +
geom_point(color = "dodgerblue") +
xlab("current shrub volume") +
ylab("total dry weight") +
geom_smooth(method = lm)
ggplotly(p)fit <- lm(total.dry.weight.g ~ volume, data= merged.data)
summary(fit)##
## Call:
## lm(formula = total.dry.weight.g ~ volume, data = merged.data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -4458 -1068 1070 1070 1070
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 1080.63 187.51 5.763 1.74e-08 ***
## volume 171.02 12.49 13.694 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1549 on 370 degrees of freedom
## (172 observations deleted due to missingness)
## Multiple R-squared: 0.3363, Adjusted R-squared: 0.3346
## F-statistic: 187.5 on 1 and 370 DF, p-value: < 2.2e-16#number of resprouted shoots predict dry weight
p4 <- ggplot(merged.data, aes(resprout, total.dry.weight.g)) +
geom_point(color = "dodgerblue") +
xlab("resprouted shoot numbers") +
ylab("total dry weight") +
geom_smooth(method = lm)
ggplotly(p4)fit <- lm(total.dry.weight.g ~ resprout, data= merged.data)
summary(fit)##
## Call:
## lm(formula = total.dry.weight.g ~ resprout, data = merged.data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -4455.8 -465.4 836.8 836.8 836.8
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -639.78 262.31 -2.439 0.0152 *
## resprout 466.79 29.02 16.087 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1459 on 370 degrees of freedom
## (172 observations deleted due to missingness)
## Multiple R-squared: 0.4116, Adjusted R-squared: 0.41
## F-statistic: 258.8 on 1 and 370 DF, p-value: < 2.2e-16merged.data <-merged.data %>% mutate(relative.dry.biomass.regrowth = total.weight/setup.vol) #calculate proportion reprouting by initial size
library(outliers)
chisq.out.test(merged.data$relative.dry.biomass.regrowth)##
## chi-squared test for outlier
##
## data: merged.data$relative.dry.biomass.regrowth
## X-squared = 1.4361, p-value = 0.2308
## alternative hypothesis: highest value 1224.12420382166 is an outliermerged.data <- merged.data %>% filter(relative.dry.biomass.regrowth < 1224)
p <- ggplot(merged.data, aes(volume, relative.dry.biomass.regrowth)) + geom_point(color = "dodgerblue") + xlab("current shrub volume") +
geom_smooth(method = lm) +
ylab("dry biomass regrowth by initial shrub size") +
scale_y_continuous(breaks=c(-2,-1,0,1,2))
ggplotly(p)m <- lm(merged.data$relative.dry.biomass.regrowth ~ merged.data$volume)
summary(m) # slope is not significantly different from 0 so no difference in proportionate regrowth##
## Call:
## lm(formula = merged.data$relative.dry.biomass.regrowth ~ merged.data$volume)
##
## Residuals:
## Min 1Q Median 3Q Max
## -42.639 -26.493 -26.493 -1.809 164.079
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 41.0911 3.5949 11.430 <2e-16 ***
## merged.data$volume 0.6456 0.3770 1.712 0.0878 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 53.25 on 319 degrees of freedom
## Multiple R-squared: 0.009109, Adjusted R-squared: 0.006003
## F-statistic: 2.933 on 1 and 319 DF, p-value: 0.08778anova(m, test = "Chisq")## Analysis of Variance Table
##
## Response: merged.data$relative.dry.biomass.regrowth
## Df Sum Sq Mean Sq F value Pr(>F)
## merged.data$volume 1 8317 8317.1 2.9326 0.08778 .
## Residuals 319 904714 2836.1
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1#Morphometrics analyses####
p <- ggplot(merged.data, aes(total.singletons, total.complex.green.shoots)) + geom_point(color = "dodgerblue") +
geom_smooth(method = lm)
ggplotly(p) p <- ggplot(merged.data, aes(resprout, total.singletons)) +
geom_point(color = "dodgerblue") +
geom_smooth(method = lm)
ggplotly(p) p <- ggplot(merged.data, aes(volume, total.singletons)) +
geom_point(color = "dodgerblue") +
geom_smooth(method = lm)
ggplotly(p)Summary
Primary trends.
Counting resprouted shoots is an effective and relatively accurate mechanism to estimate regrowth in Ephedra californica following damage. Trends in regrowth similar between shoot number and total wet weight, and shoot number significantly and effectively predicts the mass of biomass regrowth per shrub.
Larger shrubs have more resprouted shoots and regrowth of biomass; however, the regrowth is not proportionately greater for increasingly larger shrubs. THIS is an important finding. Larger shrubs do NOT have a greater capacity to proportionally recover from damage relative to smaller shrubs.
Ephedra californica can recover to some extent following damage. However, we do not know how these recovery patterns compare to the background growth patterns of undamaged shrubs nor whether subsequent or repeated damages to this species reduces its potential capacity to recover.
Methodologically, wet weight is an effective measure of biomass produced (i.e. predicts dry weight effectively with little variation).
Future directions.
Remeasure recovery in 1 and 2 years to this new clipping event.
Develop a means to estimate growth rate in undamaged shrubs for the next 1-2years.